Loop optimization
Download as PPTX, PDF3 likes6,832 views
Loop optimization is a technique to improve the performance of programs by optimizing the inner loops which take a large amount of time. Some common loop optimization methods include code motion, induction variable and strength reduction, loop invariant code motion, loop unrolling, and loop fusion. Code motion moves loop-invariant code outside the loop to avoid unnecessary computations. Induction variable and strength reduction techniques optimize computations involving induction variables. Loop invariant code motion avoids repeating computations inside loops. Loop unrolling replicates loop bodies to reduce loop control overhead. Loop fusion combines multiple nested loops to reduce the total number of iterations.
![Code Motion: -
⢠For Example: -
while( i <= max - 1 )
{
sum = sum + a[i];
}
⢠Optimized Code: -
n = max â 1;
while( i <= n )
{
sum = sum + a[i];
}](https://image.slidesharecdn.com/loopoptimization-180418113642/85/Loop-optimization-6-320.jpg)
![InductionVariable: -
⢠For Example: -
B1
i := i + 1
t1 := 4*j
t2 := a[t1]
If t2 < 10 goto B1
⢠In above code the value of i and t1 are in
locked state.
⢠That is, when value of i gets incremental
by 1 then t1 gets incremented by 4.
⢠Hence i and t1 are induction variables.](https://image.slidesharecdn.com/loopoptimization-180418113642/85/Loop-optimization-8-320.jpg)
![Loop Unrolling: -
⢠For Example: -
int i = 1;
While ( i <=100 )
{
a[i] = b[i];
i++;
}
⢠Can be written as
int i = 1;
While ( i <=100 )
{
a[i] = b[i];
i++;
a[i] = b[i];
i++;
}](https://image.slidesharecdn.com/loopoptimization-180418113642/85/Loop-optimization-12-320.jpg)
![Loop Fusion: -
⢠For Example: -
for i := 1 to n do
for j := 1 to m do
a[i, j] := 10
⢠Can be written as: -
for i := 1 to n*m do
a[i] := 10](https://image.slidesharecdn.com/loopoptimization-180418113642/85/Loop-optimization-13-320.jpg)
Loop optimization
- 1. G H Patel College of Engineering &Technology Submitted By: - GandhiVivek (140110107017) Gujarat Technological University SubmittedTo: - Prof. Namrata Dave
- 3. Code Optimization: - ⢠The code optimization can be significantly done in loops of the program. ⢠Specially inner loop is a place where program spends large amount of time. ⢠Hence if number of instruction are less in inner loop then the running time of the program will get decreased to a large extent. ⢠Hence loop optimization is a technique in which code optimization performed on inner loops.
- 4. Methods of Loop Optimization: - 1. Code Motion 2. InductionVariable and Strength Reduction 3. Loop Invariant Method 4. Loop Unrolling 5. Loop Fusion
- 5. Code Motion: - ⢠Code motion is a technique which moves the code outside the loop. ⢠If there lies some expression in the loop whose result remain unchanged even after executing the loop several times, then such an expression should be placed just before the loop.
- 6. Code Motion: - ⢠For Example: - while( i <= max - 1 ) { sum = sum + a[i]; } ⢠Optimized Code: - n = max â 1; while( i <= n ) { sum = sum + a[i]; }
- 7. InductionVariable and Strength Reduction: - ⢠A variable x is called an induction variable of loop L if the value of variable gets changed every time. ⢠It is either decremented or incremented by some constant.
- 8. InductionVariable: - ⢠For Example: - B1 i := i + 1 t1 := 4*j t2 := a[t1] If t2 < 10 goto B1 ⢠In above code the value of i and t1 are in locked state. ⢠That is, when value of i gets incremental by 1 then t1 gets incremented by 4. ⢠Hence i and t1 are induction variables.
- 9. Strength Reduction: - ⢠The strength of certain operators is higher then others. ⢠For Example: - ⢠Strength of 8 is higher than +. for ( i = 1; i <= 50; i++) count = i * 7; ⢠Here we get values of count as 7, 14 and so on. temp = 7; for ( i = 1; i <= 50; i++) { count = temp; temp = temp + 7; }
- 10. Loop Invariant Method: - ⢠In this optimization technique the computation inside the loop is avoided and there by the computation overhead on compiler is avoided. ⢠This ultimately optimizes code generation.
- 11. Loop Invariant Method: - ⢠For Example: - for i := 0 to 10 do begin K = i + (a/b); ⌠⌠end; ⢠Can be written as t := a/b; for i := 0 to 10 do begin K = i + t; ⌠⌠end;
- 12. Loop Unrolling: - ⢠For Example: - int i = 1; While ( i <=100 ) { a[i] = b[i]; i++; } ⢠Can be written as int i = 1; While ( i <=100 ) { a[i] = b[i]; i++; a[i] = b[i]; i++; }
- 13. Loop Fusion: - ⢠For Example: - for i := 1 to n do for j := 1 to m do a[i, j] := 10 ⢠Can be written as: - for i := 1 to n*m do a[i] := 10